Qiskit-aqua

Changelog

Fixed
- Don't add Hadamards to a user-defined initial state in QAOA (1362)
- Fix chemistry renos and move 0.8 renos to a subfolder (1369)
- QSVM L2 Reg. introduction for convergence and over-fitting issue (1378)
- Message enhancement in MatrixOp init (1380)
- Use PassManager in QuantumInstance transpile (1382)
- Allow Noise Simulation for simulator backends (1384)
- Fix opflow bugs (1390)
- qGAN - PyTorch Discriminator Fix (1391)
- Fix for Incorrect translation of integer variables in ADMMOptimizer (1393)
- Change logger level to debug (1395)
- Failure on PSI4 CI (1396)
- Fix ``seed_simulator=0`` in the quantum instance (1410)
- Fix code trying to iterate NoneType (1414)
- DOCplex has removed one_letter_symbol() from VarType causing problems in Aqua optimization module (1420)

Changelog
Deprecated
- Bosonic algorithms and support for Chemistry (1356)
- Deprecate the ``CircuitFactory`` and derived types (1348)
- Ground state interface (1288)
- Make QGAN run primarily on circuits (1341)
- Deprecate the ``WeightedSumOperator`` (1307)
- Properly deprecate quadratic program ising converter classes (1178)

Added
- Bosonic algorithms and support for Chemistry (1356)
- Excited States algorithms and support for Chemistry (1354)
- BOPES Sampler algorithm for Chemistry (1352)
- Orbital Optimization VQE algorithm for Chemistry (1355)
- Gradients Framework (1293)
- Implement selection of optimizer in the QGAN algorithm (1253)
- Opflow Cvar Implementation (1345)
- Ground state interface (1288)
- Finance applications (1342)
- Make QGAN run primarily on circuits (1341)
- Handle infeasible filter criterion for NumPy(Minimum)Eigensolver and MinimumEigenOptimizer (1310)
- Add support for v2 provider interface (1270)
- Create Eigensolver interface (1277)
- Optimizers for noisy quantum devices (scikit-quant) (1240)
- Vibrational structure calculations (1232)
- Operator.compose should allow indices (1144)
- Enhanced Aqgd optimizer (1252)
- HHL and NumPyLSsolver Result Classes (1254)
- Consistent behaviour of StateFn.eval with OperatorBase.eval (1210)
- More powerful Operator.to_circuit 1089 (1124)
- Add a parameters method to the Operators (1194)
- Warm start ADMM (1202)
- Fix cyclicity check and add max_iterations in VQEAdapt (1187)

Changed
- Gradients Framework (1293)
- Ground state interface (1288)
- Refactor Chemistry Drivers to also accept Molecule class (1297)
- Refactoring `Grover` class in `QuantumAlgorithm` (1268)
- Refactor QAE using the Grover operator (1234)
- Add feasibility checks to optimizers/converters (1199)
- Add Grover and QAE result classes (1219)

Removed
- Drop support for Python 3.5 (1227)

Fixed
- Setting is_measurement for converted state functions and backend arg required to CircuitSampler (1344)
- Change base gate U3 to U (1351)
- Fix measurements bug when specifying `qasm_simulator` in `GroverOptimizer` (1312)
- Fix Trotter evolution of sums with identity terms (1304)
- Fixed bugs of converters in the case of that the range of variable is 0 (1259)
- Fix a bug of the integer to binary converter (1256)
- Fix num oracle queries in QAE (1242)
- Use Operator rather than unitary simulator to convert circuit to unitary matrix (1224)
- Fix QSVM problem wit non-positive semi-definite kernel matrices (1190)
- Fix a bug in grover optimizer (1197)
- ComposedOp.eval shouldn't mutate the oplist (1207)
- correct the UCCSD.compute_excitation_lists() (1201)
- Fix cyclicity check and add max_iterations in VQEAdapt (1187)

Changelog
========

Deprecated
----------

- Properly deprecate quadratic program ising converter classes (1178)

Changelog
=========

Added
-----

- Add validation of sizes of arguments of OptimizationResults and unit tests (1167)
- For the Enhancement of converters of QuadraticProgram (1061)
- Indented string display (1112)
- Improve opflow test repeatability (1147)
- Make pyscf an optional dependency (1140)
- Documentation update for CustomCircuitOracle (1149)
- Further enhancements in ADMMOptimizer (1126)
- Check number of qubits in PauliOp.eval (1105)
- Improve inequality for "a" in Shor's Algorithm documentation (1133)
- Since 1100 Aqua depends on Terra 0.15 or higher (1119)
- For Z2Symmetry chop resultant tapered operators (1110)
- added LinearEqualityToPenalty._auto_define_penalty() function (1043)
- Unify optimizer parameters (1084)
- Reduce the number of qubits required by ADMMOptimizer (1050)
- Pretty output for optimization elements (1021)
- Make cvxpy an optional dependency (1055)
- VQE default expectation selection (1040)
- Set lstsq fitter for StateTomographyFitter on HHL (1039)
- Pin sphinx version temporarily (1033)
- Add static type hints checking (1020)
- Refactor Shor's algorithm (975)
- Re-use optimal vector result previously computed in VQE (1014)
- Update to new NumPy RNG (921)
- Allow Custom Lambda for Iterative Grover, see 893 (1005)
- Add reno for release notes and update CONTRIBUTING guide (1002)
- Replace algorithm result eigvecs by eingenstate (997)
- Improve resource access for unit tests (993)
- Speed-up of AbelianGrouper.group_subops (923)
- sort params in variational dists (985)

Changed
-------

- Move functionality QuadraticProgramToNegativeValueOracle to circuit library and GroverOptimizer (1160)
- Make OptimizationResult read-only (all parameters of the constructor become mandatory) (1131)

Removed
-------

- Remove cvxpy cap (1150)
- Remove Swaps From IQFT in Grover Optimizer (1078)
- Remove cvxopt and use cvxpy instead (1035)
- Remove install check for PySCF 1.7.2post1 (1004)


Fixed
-----

- Mp2info incorrect when only one double excitation (1161)
- Fix PySCF driver exception (1159)
- show full grover circ; fixes 1139 (1141)
- VQE change backends (1154)
- Fix preservation of ListOp's combofn (1152)
- VQE construct circuit returns circuits (1121)
- Fix torch abstract class method not implemented (1146)
- OptimizationResult does not contain variable information 1107 (1122)
- Be more memory efficient converting SummedOp to MatrixOp (1129)
- Matrix multiplicative factor (z + z == 2 * z) (1111)
- Determine the max number of ancillae for all ESOPs (TruthTableOracle fix) (1128)
- P_BFGS spawn crash on MacOS and python >= 3.8 (1123)
- Fix qgan.py where batch size is greater than data size to raise error (1115)
- Fix to accommodate latest docplex 2.15.194 (1117)
- SLSQP optimizer, multistart Cobyla, enhanced compatibility checks (1082)
- Fix lint in controlled circuit and Instruction.definition change (1100)
- SummedOp updates & optimization converters to use Opflow (1059)
- CircuitStateFn.assign_parameters overwrites is_measurement (1094)
- OperatorStateFn: bind parameters in coeff and primitive (1085)
- Set qcp to True for cvxpy Problem.solve in qp_solver (1083)
- Z2 symmetry fix for when no symmetries found (1075)
- DeprecationWarning of round with numpy 1.19 (1081)
- Correct coefficient propagation in the operators (1077)
- Fix Exchange/Wiki providers, new Yahoo! provider, fix divide by 0 (1051)
- remove warnings from deprecated varform (1058)
- Fix hash function of PauliOp class (1052)
- MatrixOp.to_instruction updated to return an Instruction (1049)
- Evaluate eigenvalues in MinEigenOptimizer more efficiently (1045)
- CircuitOp uses QuantumCircuit.compose, not .combine (1048)
- fix remaining type hints errors (1025)
- Fix QAOA initial point bug (1026)
- Remove deprecated VQ tests (987)
- Fix ADAM resuseability (1016)
- Test lam argument in Grover algorithm (1011)
- Fix VQE optimizer setter (1008)
- Separate simplification of operators from SummedOp.add and fix bugs of add (1000)
- Fix cobyla optimizer (999)
- Ignore buggy pyscf 1.7.2.post1 (994)

Changelog
=========

Fixed
-----

- CircuitOp uses QuantumCircuit.compose, not .combine (1048)
- MatrixOp.to_instruction updated to return an Instruction (1049)
- Fix hash function of PauliOp class (1052)
- Make cvxpy an optional dependency (1055)

Changelog
=========

Fixed
-----

- Remove cvxopt and use cvxpy instead (1036)
- Pin sphinx version temporarily (1038)
- Fix VQE optimizer setter, VQE default expectation selection (1046)